CFD Online Discussion Forums

CFD Online Discussion Forums (
-   OpenFOAM Running, Solving & CFD (
-   -   Problem with simpleFoam (

Daniele111 July 6, 2010 03:33

Problem with simpleFoam
When I run my case with simpleFoam with a refine mesh, after some iteration I have this error message why? (my k and epsilon aren't zero)

time step continuity errors : sum local = 2.25645e+45, global = -2.2833e+38, cumulative = -2.2833e+38
#0 Foam::error:printStack(Foam::Ostream&) in "/user/parolini/OpenFOAM/OpenFOAM-1.7.x/lib/linux64GccDPOpt/"
#1 Foam::sigFpe::sigFpeHandler(int) in "/user/OpenFOAM/OpenFOAM-1.7.x/lib/linux64GccDPOpt/"
#2 __restore_rt at sigaction.c:0
#3 Foam::PBiCG::solve(Foam::Field<double>&, Foam::Field<double> const&, unsigned char) const in "/user/OpenFOAM/OpenFOAM-1.7.x/lib/linux64GccDPOpt/"
#4 Foam::fvMatrix<double>::solve(Foam::dictionary const&) in "/user/OpenFOAM/OpenFOAM-1.7.x/lib/linux64GccDPOpt/"
#5 Foam::lduMatrix::solverPerformance Foam::solve<double>(Foam::tmp<Foam::fvMatrix<doubl e> > const&) in "/user/OpenFOAM/OpenFOAM-1.7.x/lib/linux64GccDPOpt/libinc$
#6 Foam::incompressible::RASModels::kEpsilon::correct () in "/user/OpenFOAM/OpenFOAM-1.7.x/lib/linux64GccDPOpt/"
#7 main in "/user/OpenFOAM/OpenFOAM-1.7.x/applications/bin/linux64GccDPOpt/simpleFoam"
#8 __libc_start_main in "/lib64/"
#9 Foam::regIOobject::writeObject(Foam::IOstream::str eamFormat, Foam::IOstream::versionNumber, Foam::IOstream::compressionType) const in "/user/OpenFOAM/OpenFOAM$

Mo-ITB July 6, 2010 04:56

hi daniele,

time step continuity errors : sum local = 2.25645e+45

this seems to produce a floating point exception caused by too large numbers.

maybe check the velocity field of the last time step before divergence?
could be that a bc is not set correctly.


Daniele111 July 8, 2010 08:06

1 Attachment(s)
Time step error explode when I refine mesh, i don't understand...
this is my case, is a atmospeheric bundary layer with a obstacle on the bottom

Mo-ITB July 8, 2010 08:10

have you tried to change the relaxation factors for p and U?
and maybe initialize with some magnitude!

Daniele111 July 8, 2010 08:15

Initialize p and U in 0 dir?. U is initialized but p is set equal zero, I try to change it . How can change relaxation factor?

Thank you

Mo-ITB July 8, 2010 08:27

no, initializing U is enough ;). the relaxation factors are set in system/fvSolution like that:

U 0.3;
p 0.7;

they sould be 1 in sum. for the beginning 0.3 for U may be better, you might change it to 0.7 and p to 0.3 after a while.

Daniele111 July 8, 2010 08:31

And factors for k epsilon? 0,8 is good? What value can i use to initialize U?


Mo-ITB July 8, 2010 08:40

you can try to switch off turbulence in constant/RASProperties and calculate some laminar iterations. then you can start from there with turbulence switched on. i don't know which factors for k and eps to use, depends on your case, but you can start with 0.8 and decrease them if its diverging.

Daniele111 July 8, 2010 09:04

1 Attachment(s)
With the factor that you give me time step continuity error is ok! Thanks! You are very kind. My result now are better. Can I ask you another thing? This is my wall shear stress on bottom The obstacle has a step with a ricircularion bubble; and in my plot I have a overshooting. How can i try to eliminate it?

Mo-ITB July 8, 2010 09:43

sorry i have no experience with that yet... maybe someone else :).

All times are GMT -4. The time now is 16:17.